US-12625110-B2 - Complex part inspection with eddy current sensors

US12625110B2US 12625110 B2US12625110 B2US 12625110B2US-12625110-B2

Abstract

Eddy current sensing is governed by the diffusion equation of magnetoquasistatic fields. As such the eddy current sensor's proximity to the object to be inspected (i.e., “liftoff”) significantly affects the sensor's response signal. Methods and apparatus are disclosed for improving performance for an eddy current sensor, though they may also be used for other sensor types. These solutions are beneficial for both single channel eddy current sensors and arrays, and are particularly beneficial for measuring parts with complex surfaces. In some aspects improved performance is achieved by varying the stiffness of the mechanical support for the sensor. Some mechanical supports may exhibit anisotropic stiffness. After performing a scan with an eddy current array, a multi-channel shape filtering module is applied to improve defect detection. The module reduces the variability of defect response measured due to the unpredictability of the defect location transverse to the scan direction.

Inventors

Neil J Goldfine
Mark Windoloski
Todd M Dunford

Assignees

JENTEK SENSORS, INC.

Dates

Publication Date: 20260512
Application Date: 20220905

Claims (18)

1 . A system comprising: at least one processor; a sensor array having a plurality of sense elements; an encoder to record a spatial position of the sensor array; a memory unit that is at least one non-transitory computer-readable storage medium, the memory unit having stored thereon a signature library, the signature library comprising a plurality of defect signatures, each defect signature having, on a first number of channels, a defect response as a function of spatial position, the first number being two or more, and at least one software module comprising instructions executable by the at least one processor; an instrument configured to collect measurements from each of the plurality of sense elements of the sensor array as a function of the spatial position obtained from the encoder; and a correlation module to correlate, as a function of spatial position, the measurements from a subset of sense elements with each defect signature in the signature library, wherein the subset of sense elements has the first number of sense elements, wherein the correlation module is a software module among the at least one software module, wherein the correlation module generates a correlation value, a detection module to determine a defect is present when the correlation value exceeds a threshold, and an output module to output a signal indicating a presence of the defect.
2 . The system of claim 1 , wherein the subset of sense elements are adjacent sense elements among the plurality of sense elements in the sensor array.
3 . The system of claim 1 , wherein the plurality of sense elements consists of a second number of sense elements, and the correlation module performs the correlation for a third number of subsets of sense elements, wherein the third number is equal to the second number minus the first number plus one.
4 . The system of claim 3 , wherein the correlation module is further configured to determine a maximum correlation value from among a first number of correlations determined for each spatial position for each subset and store the maximum correlation value in the memory unit, and the system further comprises a display configured to output a visual representation of the maximum correlation values.
5 . The system of claim 1 , wherein the correlation module is further configured to determine a maximum correlation value from among a first number of correlations determined for each encoder position and store the maximum correlation value in the memory unit.
6 . The system of claim 1 , wherein each of the plurality of defect signatures has a same spatial length and each channel of each signature has zero mean, the at least one software module further comprising a measurement preprocessing module configured (i) to resample the measurements to have a measurement spacing equal to that of the defect signatures; (ii) to define measurement sets from the measurements to be correlated by the correlation module with each measurement set having the spatial length; and (iii) to remove from each measurement set its mean so that each measurement set has zero mean.
7 . The system of claim 1 , further comprising a multivariate inverse method module that applies a model based inverse method to the measurements to estimate a material property to be correlated with the signatures of the signature library.
8 . The system of claim 1 , wherein the sensor array is an eddy current sensor array having a common drive winding shared by the plurality of sensing elements.
9 . The system of claim 1 , wherein the sensor array is a capacitive sensor array having one or more drive electrodes and where each sense element is a sense electrode.
10 . The system of claim 1 , wherein the plurality of sense elements consists of a second number of sense elements, and the instrument collects measurements from the second number of sense elements using a second number of parallel measurement channels.
11 . The system of claim 10 , wherein each of the parallel measurement channels of the instrument simultaneously measures a real part and an imaginary part of the respective measurement.
12 . They system of claim 1 , wherein the signature library stored in the memory unit comprises defect signatures obtained at a plurality of liftoffs.
13 . They system of claim 1 , wherein the signature library is a first signature library for a first frequency, the memory unit stores a second signature library having defect signatures obtained at a second frequency, the instrument collects measurements at the first and second frequencies, and the correlation module correlates measurements at the first frequency with the defect signatures of the first signature library and correlates measurements at the second frequency with the defect signatures of the second signature library.
14 . A method of defect detection, the method comprising: scanning a sensor array having a plurality of sense elements over a test object; during the scanning recording a spatial position of the sensor array; collecting measurements from each of the plurality of sense elements of the sensor array as a function of the spatial position; and operating a processor to correlate as a function of spatial position, the measurements from a subset of sense elements with each defect signature in a signature library, wherein each defect signature in the signature library comprises, on a first number of channels, a defect response as a function of spatial position, the first number being two or more, and the subset of sense elements has the first number of sense elements, wherein the processor generates a correlation value based on the correlating, the processor determines a defect is present when the correlation value exceeds a threshold, and the processor outputs a signal indicating a presence of the defect.
15 . The method of claim 14 , wherein each of the defect signatures in the signature library has a same spatial length and each channel of each signature has zero mean, the method further comprising resampling the measurements to have a measurement spacing equal to that of the defect signatures; defining measurement sets from the measurements to be correlated by the correlation module with each measurement set having the spatial length; and removing from each measurement set its mean so that each measurement set has zero mean.
16 . The method of claim 14 , wherein the signature library comprises a first defect signature obtained at a first liftoff and a second defect signature obtained at a second liftoff.
17 . The method of claim 14 , wherein the subset of sense elements are adjacent sense elements among the plurality of sense elements in the sensor array.
18 . A system for defect detection, the system comprising: at least one processor; a sensor array having a plurality of sense elements; an encoder to record a spatial position of the sensor array; an instrument configured to collect measurements from each of the plurality of sense elements of the sensor array as a function of the spatial position obtained from the encoder; and at least one non-transitory computer-readable storage medium having stored thereon a signature library, the signature library comprising a plurality of defect signatures obtained at a plurality of liftoffs, each defect signature having, on a first number of channels, a defect response as a function of spatial position, the first number being two or more, and a plurality of software modules, each comprising instructions executable by the at least one processor, including a correlation module to correlate, as a function of spatial position, the measurements from a subset of sense elements with each defect signature in the signature library, and generate a correlation value, wherein the subset of sense elements has the first number of sense elements, a detection module to determine a defect is present when the correlation value exceeds a threshold, and an output module to output a signal indicating a presence of the defect.

Description

RELATED APPLICATION The present application is a continuation is U.S. application Ser. No. 17/046,047 filed Oct. 8, 2020, which claims the benefit of International Application No. PCT/2019/026673 with an international filing date of Apr. 9, 2019, which itself claims priority to U.S. provisional patent application, U.S. Ser. No. 62/654,691, filed Apr. 9, 2018, which are herein incorporated by reference in their entirety. TECHNICAL FIELD The present disclosure relates to the field of measurement apparatus and methods. Some aspects of the disclosure relate to eddy current sensing and eddy current arrays. BACKGROUND ART Eddy current is a electromagnetic phenomenon that has been utilized to inspect materials for cracks, corrosion, porosity, and may other defect types. Eddy current sensing is governed by the diffusion equation of magnetoquasistatic fields. As such, the eddy current sensor's proximity to the object to be inspected (i.e., “liftoff”) significantly affects the sensor's response signal. The defect detection capability thus becomes unacceptable at larger liftoffs. Practical applications exists where liftoffs above just several thousandths of an inch result in unacceptable defect detection performance. Eddy current sensors can be built with one or more measurement channels. Eddy current arrays have a number of sense elements. The defect signal produced by eddy current arrays varies with the transverse position of the sense elements of the array relative to the defect. If a sense element passes directly over the defect the element generally produces a larger response than if the element is not centered on the defect. Generally the smallest element responses are observed when the array passes over the defect with the defect falling directly between two sense elements. SUMMARY Methods and apparatus are disclosed for improving performance for an eddy current sensor. The sensor may be mounted to a mechanical support that provides variable stiffness along the surface. By varying the stiffness the sensor may better conform to the surface of a complex part to be inspected. Improving the conformance of the sensor to the curvature of the part reduces the liftoff of the sensor. Under most conditions reduced liftoff improves the defect detection performance of the sensor. The stiffness of the mechanical support can be varied by including captured volumes within the mechanical support. These volumes may simply be void (typically air filled) or filled with another material of a different stiffness. By varying the size, location, fill material, and spacing of these captured volumes the stiffness observed at the sensor mounting surface can be varied in ways that improve conformance to a complex part. The mounted sensor may be scanned along a complex part while measurement data is collected. The variable stiffness at the sensor surface may allow the sensor to maintain conformity with the complex part even if the curvature of the part varies along the scan path. The sensor measurements can be processed to enhance observability of defects. Multivariate inverse methods based on physics based models may be used to isolate material response from the sensor response due solely to liftoff. The material response can be enhanced by accounting for the transverse position of the sense elements relative to any defects in the inspected part. Generally the transverse location of the defect is not known apriori. A signature library is used to process the measurement data from multiple adjacent sense elements (or equivalently adjacent passes of the sensor. The signature library contains a set of multi-channel defect signatures. Each defect signature can be made from the measured response of selected measurement channels when the sensor passes over the defect at different relative transverse locations. The measurement data is correlated with each of the signatures. Assuming a defect is present, the correlation will generally be largest with the signature having the closest transverse defect position as the measurement data. Advantageously a given defect will have approximately the same maximum correlation value regardless of the transverse position of the defect in a given scan. Some aspects relate to a sensor cartridge comprising a sensor and a mechanical support. The mechanical support has a surface to which the sensor is secured. A first component of the mechanical support may be made with a first material and may have a plurality of captured volumes. The captured volumes are devoid of the first material. The surface to which the sensor is secured may thus have variable stiffness which improves conformity of the sensor cartridge to a feature of a test object when the sensor cartridge is pressed against the feature. The capture volumes may be holes, pits, internal pores, and the like. The captured volumes may be empty or filled with a material having a different stiffness than the first material. In the case of holes, the holes may ea